| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Yonsei University College of Medicine, Seodaemun-ku
Shinchon-dong 134, Seoul 120-752, South Korea.
2 Department of General Surgery, Yonsei University College of Medicine, Seoul,
South Korea.
3 Department of Oncology, Yonsei University College of Medicine, Seoul, South
Korea.
Received March 31, 2008;
accepted after revision July 17, 2008.
Address correspondence to E. K. Kim
(ekkim{at}yuhs.ac).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Between January 2003 and December 2003, 1,706 breast sonographic examinations were performed by three radiologists in 1,256 Asian women with a history of surgery for breast cancer in one breast as an adjunct screening test to mammography in an academic medical center. We evaluated the biopsy recommendation rate, a diagnostic index, of the combination of whole-breast sonography and mammography for the detection of contralateral metachronous breast cancers and the positive predictive value (PPV) of this biopsy recommendation rate.
RESULTS. Based on 1,706 examinations in 1,256 women, the biopsy recommendation rate was 3.5% per patient and 2.6% per examination. The PPV of the biopsy recommendation rate was 41.0% with 18 breast cancers diagnosed (cancer detection rate, 1.4% per patient and 1.1% per examination). Among these cancers, two were detected on sonography alone. One false-negative cancer was found on the next sonographic examination but could not be seen on the next mammographic examination.
CONCLUSION. With a false-negative rate of only 0.06% and a PPV of 41.0% for the biopsy recommendation rate, our results suggest that annual sonography could be a useful adjunctive tool to mammography for the detection of metachronous contralateral cancers.
Keywords: breast cancer • breast cancer screening • breast sonography • mammography • women's imaging
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Here, we assessed the diagnostic indexes of sonographic surveillance for the detection of metachronous contralateral breast cancer in consecutive patients with a history of breast cancer surgery.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Study Population
Between January 2003 and December 2003, 1,901 adjunctive sonographic
examinations were performed in 1,431 patients who had undergone a mastectomy
or breast-conservation surgery for breast cancer at our institution. All of
the patients who were 40 years old or older had an annual mammogram that
showed grade 2-4 dense parenchymal tissues
[3]. When a patient younger
than 40 years had suspicious findings on sonography, mammography was
recommended. Among the 1,901 sonographic examinations, we excluded 39
examinations in 28 patients because of a history of operation for bilateral
breast cancers. These 39 examinations were performed to evaluate the chest
wall and remaining breast tissue after ipsilateral breast-conserving surgery.
The remaining 1,862 screening sonograms in 1,403 patients included the
follow-up examinations of the contralateral breast in patients without any
history of contralateral breast cancer surgery. All the women were Asian.
Among these examinations, 131 examinations in 123 patients were excluded because the patient was lost to follow-up within 1 year. Twenty-five examinations in 24 patients were also excluded because distant metastatic foci were detected: These individuals did not undergo any additional screening study for contralateral breast cancer. Finally, of the 1,706 sonographic examinations in 1,256 patients comprising our study population, 806 patients underwent sonography once and 450 patients, twice. Of these, 1,187 patients (94.5%) who underwent 1,619 examinations were followed for at least 2 years. The mean interval time from previous surgery for breast cancer to the time when the examinations were performed in the study period was 39 months (range, 6-252 months).
Surveillance
Since 2000, all patients at our institution with a history of breast cancer
surgery except those with entirely fatty breasts on mammography have been
advised to undergo bilateral whole-breast sonography as well as mammography
[1,
2]. Clinical follow-up
examinations after breast cancer surgery were performed every 6 months for the
first 2 or 3 years after surgery and then annually thereafter. During clinical
follow-up examinations, patients with invasive breast cancer were advised to
undergo annual mammography and bilateral whole-breast sonography every 6
months for the first 2 years and then annual mammographic and sonographic
evaluations thereafter. Patients with ductal carcinoma in situ were advised to
undergo annual mammographic and sonographic evaluations. In patients who had
undergone a mastectomy for breast cancer, sonographic evaluations of the chest
wall and contralateral breast were performed. The interval between imaging
studies varied somewhat because of patient preference and examination
scheduling. Some of our patients could not accept our recommended protocol
because of their schedules or did not accept our protocol.
Imaging Evaluation
Mammograms were obtained with dedicated equipment (DMR, GE Healthcare).
Standard craniocaudal and mediolateral oblique views were routinely obtained
and additional mammographic views were obtained as needed. Bilateral
whole-breast sonography has been prospectively performed in our institution
since 2000. Each sonographic examination was performed using either an ATL HDI
5000 or 3000 sonography unit (Philips Healthcare) with 5-10-MHz or 5-12-MHz
linear-array transducers by three full-time board-certified radiologists. At
the time of the study, the three board-certified radiologists had 4, 4, and 10
years' experience, respectively, in performing breast sonography and
interpreting breast sonography and mammography at an academic medical center.
The number of screening breast sonographic examinations performed by the
radiologists in their practice varied from 200 to 350 examinations per month.
The radiologist performing sonography knew the results of each preceding
examination before performing the succeeding examination as well as the
results of previous screening studies, if any, 1 year or 6 months earlier.
In our institution, mammography is performed the same day that sonography is scheduled, but mammography is performed before sonography. The mammograms were reviewed by the radiologist who performed sonography at the time of the examination, and sonography was performed without an officially reported interpretation. If mammography was not planned at the same time as sonography, sonography was performed first and if an abnormality was detected on sonography, mammography was recommended before biopsy.
Sonographic examinations were performed with patients in the supine position with their arms stretched over their heads. When necessary, patients were shifted to the contralateral posterior oblique position so that the lateral and inferior aspects of the breast could be scanned. Scanning was targeted first to the ipsilateral breast or chest wall and then to the contralateral breast during each examination, and scanning was performed in radial and antiradial planes, longitudinal and transverse planes, or both [2, 4]. Each examination took approximately 10 minutes (range, 5-20 minutes). If an abnormal lesion suspicious for malignancy was seen on sonography alone or on both sonography and mammography, a sonographically guided core needle biopsy was performed immediately with informed consent from the patient.
Sonographically guided core needle biopsies were performed using a disposable automated 14-gauge biopsy needle with a 22-mm throw (Monopty, Bard Peripheral Technologies). We used a standard freehand technique for all sonographically guided biopsies of breast lesions. Five core samples were routinely obtained. If the mammographic abnormality suspicious for malignancy was not delineated on sonography, such as microcalcifications, an excisional biopsy with mammographically guided localization was scheduled. (Stereotactic equipment was not available during the study period, but now it is available in our institution.)
Interpretation of the Contralateral Breast Images and Establishment of Final Diagnosis
One radiologist reviewed the record of imaging studies to determine BI-RADS
category classification. The BI-RADS categories of mammographic and
sonographic findings were classified according to BI-RADS category of the
original radiologic reports and were not reread to rule out the possibility
that foreknowledge of the correct cancer diagnosis may affect reinterpretation
[3]. These conditions were used
to ensure that the results would precisely reflect the accuracy of routine
diagnostic work.
According to the BI-RADS category recommendation, one integrated final assessment should consider all breast imaging findings including bilateral breast or contralateral breast and chest wall findings [3]. Assignment of findings to BIRADS category 1 or 2 as a final assessment on the original imaging report was considered a negative result for the contralateral breast. In cases in which BI-RADS category 3, 4, or 5 was assigned on the original report, a radiologist compared the original report with imaging findings and then determined whether the final assessment corresponded to the lesions in the contralateral breast or ipsilateral breast with a prior cancer operation. If the final assessment was for the lesion in the contralateral breast, the assessment was considered as that for the contralateral breast. If the assessment was for the lesion in the ipsilateral breast with a surgical history, a new final assessment was adapted for the lesions limited to the contralateral breast by the radiologist according to the original report.
Although the sonography BI-RADS lexicon did not exist before late 2003, sonographic interpretations in our institution were categorized according to the risk of malignancy like mammographic BI-RADS since 1999. In early 2003, we subclassified category 4 into categories 4a and 4b, and we regarded category 4b as equivalent to category 4c of the published BI-RADS lexicons with respect to the risk of malignancy. In late 2003, we subclassified category 4 into categories 4a, 4b, and 4c. However, the category 4b since late 2003 would have mostly been classified as 4a before the new edition of the BI-RADS lexicon was published. Because we did not have a pure 4a group, we classified our cases into category 4a or 4b and category 4c for data analysis.
The final diagnosis of each patient was based on the tissue diagnosis and at least 1 year of follow-up imaging studies. In cases in which long-term follow-up data for at least 2 years were available, we reviewed the follow-up data to determine whether there was any additional false-negative diagnosis in our study.
Data Analysis
We evaluated clinical, imaging, and pathologic findings of the patients
included in our study to assess the method of detecting metachronous cancers.
On the basis of the data obtained from the 1,706 sonographic examinations in
1,256 patients with a validated sonography diagnosis, diagnostic indexes
including the biopsy recommendation rate and the true-positive (TP),
true-negative (TN), false-positive (FP), and false-negative
(FN) rates were determined. The sensitivity, specificity, positive
predictive value (PPV), and cancer detection rate were also determined.
Sensitivity was defined as the percentage of cancers detected among all
cancers detected with any technique at the 1-year follow-up:
 |
 |
 |
 |
Patients were grouped according to whether a prior screening sonographic examination had been performed within 12 months of the screening sonographic examination in the period included for our study: Group A included patients who had undergone screening sonography and mammography within 12 months and group B, those who had not. The categories and diagnostic indexes were also determined and compared for each group. Methods for detecting metachronous cancers were compared between the two groups.
Statistical comparisons were performed using a chi-square or Fisher's exact test for nonparametric variables. Statistical significance was assigned to p values less than 0.05. Data were analyzed using an SPSS software package (version 12.0, SPSS).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Among the 1,706 sonographic examinations of 1,256 patients, the imaging reports assigned BI-RADS categories as follows (Table 2): category 1 or 2 in 1,597 (93.6%), category 3 in 65 (3.8%), category 4a or 4b in 35 (2.1%), category 4c in four (0.2%), and category 5 in five (0.3%). Biopsy was recommended for all patients with category 4 or 5 lesions (44 examinations; 2.6% per examination and 3.5% per patient). Of those, biopsies were performed in 43 lesions. An additional three biopsies for category 3 lesions were also performed because of patient anxiety. The biopsy rate was 2.7% of sonograms (46 of 1,706 sonograms) and 3.7% of patients (46 of 1,256 patients). Sonographically guided biopsy was performed in 39 patients and excisional biopsy with mammographically guided localization was performed in the remaining seven patients.
|
Follow-Up Results and Diagnostic Indexes
Eighteen malignancies (true-positive results) were found in the 46 biopsies
performed (Table 3). The number
of malignancies per BI-RADS category were as follows: no malignancies in
category 3 lesions (0.0%), nine malignancies in category 4a or 4b (25.7%),
four malignancies in category 4c (100.0%), and five malignancies in category 5
(100.0%, Table 2). The PPV of
biopsy recommended was 41.0% (18 of 44 category 4 and 5 lesions).
|
Except the 18 sonograms that revealed malignancy, the 26 examinations recommended on the basis of both sonograms and mammograms were false-positives and accounted for 1.5% of the examinations performed (n = 1,706) and for 2.1% of the patients included (n = 1,256). Among the 26 false-positive examinations, 25 pathologic results were as follows: 11 cases of fibrocystic change; five, stromal fibrosis; three, ductal ectasia; two, epithelial hyperplasia; one, intraductal papilloma; one, adenosis; one, lactating adenoma; and one, atypical ductal hyperplasia. In the remaining case, one category 4a lesion on sonography was not biopsied but showed no interval change in size on the 2-year follow-up sonogram. In the case of atypical ductal hyperplasia, a subsequent surgical excision was performed, and the pathologic result was atypical ductal hyperplasia without evidence of carcinoma.
One cancer was missed, a false-negative result, and accounted for 0.06% of the examinations performed (n = 1,706) and for 0.08% of the patients included (n = 1,256) within the 1-year follow-up. The missed cancer was a 7-mm invasive ductal carcinoma detected as a category 4c lesion on the 11-month follow-up screening sonogram; this tumor was not seen on mammography when it was diagnosed on sonography. The remaining 1,661 sonographic examinations were free of cancer at 1-year follow-up and were considered to be true-negatives.
The sensitivity of mammography and sonography together was 94.7% (18 of 19 malignancies) and the specificity was 98.5% (1,661 of 1,687 negative examinations). Sonography alone found 11.1% (two of 18) of all cancers detected, 7.1% (one of 14) of all invasive cancers, and 16.7% (two of 12) of all nonpalpable cancers. The methods of detecting metachronous breast cancers are summarized and the pathologic stages, determined after subsequent operations, are shown with the method for cancer detection in Table 4. All of the nonpalpable cancers detected on sonography only (n = 2) or mammography only (n = 3) were stage 0 or I.
|
The overall cancer detection rate was 1.4% per patient (18 of 1,256 patients) and 1.1% per examination (18 of 1,706 examinations).
Group A Versus Group B
Of the 1,256 women in the study, 761 patients (1,039 examinations) were in
group A and 495 patients (667 examinations) were in group B. Of the 12
patients who complained of palpable lumps in their contralateral breast, four
patients were in group A and eight patients were in group B.
Table 5 shows the distribution
of BI-RADS category classifications in each group. However, the biopsy
recommendation rate did not show a statistically significant difference
between the two groups (p > 0.05): 2.2% of sonograms in group A
(n = 23) versus 3.1% of sonograms in group B (n = 21).
|
The diagnostic indexes of each group showed that the PPV was higher in group B (p < 0.05) (Table 6). The methods for detecting metachronous breast cancers in each group are summarized in Table 7 (p > 0.05).
|
|
Long-Term Follow-Up
Among the 1,187 patients who underwent follow-up imaging for at least 2
years at our institution, an additional five cancers were found within 2 years
of follow-up. Of these, four were from 720 patients in group A (0.6%) and the
remaining one was from one of 467 patients in group B (0.2%; p >
0.05). Four of the five cancers were nonpalpable and of these nonpalpable
lesions, two were not detected on mammography. Four cancers were stage 0 or I.
In the remaining patient, the screening mammograms and sonograms obtained 1
year later were also negative. However, 5 months after that screening, the
patient complained of a rapidly growing lump in her breast. At that point,
mammography revealed a 2-cm hyperdense mass and sonography revealed an
irregular hypoechoic mass. A core needle biopsy was performed on the mass,
revealing highgrade ductal carcinoma. The mammograms and sonograms obtained
when the patient underwent 1-year screening were reviewed again to determine
whether the lesion was retrospectively visible; however, we could not find any
abnormality on the original screening images at the site where the patient
complained of a palpable lesion 5 months later.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Evaluation of the false-positive biopsy rate and the PPV of biopsy by sonography is mandatory for valid sonographic surveillance. If the false-positive rate is unacceptably high, then the technique cannot be practical. An excessively high false-positive biopsy rate may result in patient anxiety, recall of patients for additional evaluation, unnecessary biopsies, inconvenience, useless and unnecessary treatment, and time away from work or home. In our study, sonographic surveillance with mammography showed 94.7% sensitivity, and the biopsy recommendation rate was 2.6% per examination. The PPV of biopsies recommended was 41.0%, which is much higher than the PPV of sonographic screening results in dense breasts previously reported in the literature (11-19%) [7, 8], is also higher than the PPV according to the recommendation for mammographic screening (25-40%) [3, 12], and is similar to the PPV of sonographic screening in high-risk patients (40%; four cancers in 10 patients) reported by Crystal et al. [7]. This high PPV for biopsies recommended on the basis of the biopsy recommendation rate could suggest that a lower false-positive biopsy rate results from consideration of not only sonography but also mammography. According to the BIRADS recommendations, the final assessments should be determined prospectively on the basis of both sonographic and mammographic findings.
Determining which imaging method contributed the false-positive results is difficult because sonograms may have contributed to some cases, mammograms to other cases, and both sonograms and mammograms to the remaining cases. Although the person interpreting the sonograms cannot help but be influenced by the findings on mammography, we could have potentially contrary cases excluded from false-positive results that had been considered as positive on mammography but were determined as negative or benign on sonography. Mean while, the high PPV of the biopsy recommendation rate and the low false-positive biopsy rate may suggest that sonographic surveillance could decrease the number of false-positive cases by mammography and increase the malignancy yield rate in addition to detecting mammographically occult cancers in dense breasts. The reason our results showed a high PPV of biopsy recommendation could also be that the study population consisted of patients who are at high risk for breast cancer compared with the general population. Crystal et al. [7] also reported a higher PPV in high-risk patients than in average-risk patients (40% vs 10.7%, respectively). Because they did not comment in their work about the history of the sonography studies in the study population, we cannot determine the percentage of that population who had undergone sonography before the study period.
In our study, the PPV (57.1%) of group B was much higher than the PPV in the high-risk patients reported by Crystal et al. [7] (40%). Another possible reason for this difference in PPV results could be that most of our population had undergone breast sonography since the time of their initial breast cancer diagnosis [1, 2]. The lesions that were detected in the previous studies were classified as BI-RADS category 2 or 3 in our study period, and the lesions newly discovered and classified as category 4 or 5 in the study period could be incident cases. In a recently published article, Berg et al. [13] reported that the PPV with combined screening with sonography and mammography was 10.1% in their study population, of whom only 10% had undergone prior screening sonography. Those false-positive cases could be classified as category 2 or 3 in the next screening round as they were in our study.
In our study, the false-negative rate was an acceptable 0.06%, and the one false-negative case was detected on subsequent sonography screening 11 months later. A sonographically guided biopsy was performed soon after the abnormality was detected on sonography. Among the six cancers detected during the 1-year (n = 1) or long-term (n = 5) follow-ups, five cancers (83.3%) were nonpalpable and three were not visible on mammograms when they were diagnosed. Only one cancer might have been detected with palpation, and at most three cancers might have been detected on either mammography or palpation during long-term follow-up. These results further support the usefulness and appropriateness of sonographic surveillance. Most cancers detected during the follow-up period may have just entered that detectable phase [14].
The overall cancer detection rate was 1.4% per patient and 1.1% per examination, and the cancer detection rate by sonography alone was 0.2% per patient and 0.1% per examination. These rates are similar to those of the several previous studies that used sonographic screening [15-19] and are lower than those of other studies that used sonographic screening (0.3-0.5%) [7-9]. This discrepancy could result from the lower incidence of breast cancer in the Asian population than in the United States population and the previously performed sonography screening in our study population. Breast cancer incidence rates vary between countries, with higher rates in westernized countries and lower rates in Asian countries [20-22].
Sixty-one percent of our study population (761 of 1,256 patients) underwent sonographic surveillance within 1 year and the remaining 39% of our study population may or may not have undergone sonography screening 1 year earlier or when the initial diagnosis of breast cancer was made. In our institution since 2000, bilateral whole-breast sonography has been prospectively performed in the preoperative evaluation of patients with a known breast malignancy or in the diagnostic evaluation of patients with a suspicion of breast cancer [1]. Therefore, among most of our patients, the sonographic screening examination during our study period was not "the first round" in which many more cancers are well known to be found than would be found at subsequent screening examinations [14]. Patients with meta chronous contralateral cancers detected by sonographic surveillance who had undergone breast cancer surgery for contralateral breast cancer before our study period were excluded from our study population.
In addition to determining which technique should be used for surveillance, we also wanted to determine how often surveillance should be performed. In a previous study, investigators reported that no differences in staging were found between patients who had undergone intensive sonography within 6 months and patients who had undergone sonography every 12 months [2]. When we extended the follow-up period to 2 years, there was a 0.4% false-negative rate with no statistically significant differences between the screened and unscreened groups. From these data, we concluded that previous sonographic surveillance of the screened group did not affect the 2-year follow-up surveillance. It is possible that the time interval between 1 year and 2 years would be acceptable for sonographic surveillance, but further study should be pursued with a large study population.
Annual MRI is recommended in addition to mammography to women at very high risk of breast cancer in the United States [23]. In addition, the combined sensitivity of mammography and sonography aver aged 55% versus 93% after combined mammography and MRI across four series in which screening mammography, sonography, and MRI were performed in women with various degrees of risk for breast cancer [24-27]. We do not suggest that sonography is superior to MRI as an adjunct to mammography. However, the criteria of MRI for follow-up and risk of malignancy in lesions followed have not been well established. The use of MRI remains limited by high cost, requirement of contrast injection, reduced patient tolerance, and limited availability and operator expertise. We suggest that sonographic surveillance for an appropriate population could be more useful than we now expect it is. Berg et al. [13] suggested that sonography and MRI can be used according to the degree of risk for breast cancer. Moreover, the sensitivity of MRI depends on the degree of background parenchymal enhancement, just as the sensitivity of mammography can be limited by the high density of surrounding tissue. Half of our study population was in their fourth decade. They had dense fibroglandular tissue and the parenchymal enhancement in their breasts could yield false-negative and false-positive diagnoses, although MRI was scheduled according to the menstrual cycle. Breast sonography could be a candidate to complement this limitation of MRI.
There were some limitations to our study. First, our study population was small. Second, we did not include cost-effectiveness and survival benefit analyses; sonography screening is not covered by many insurance providers. Moreover, our study population included 12 patients who complained of a palpable lesion. We present outcomes of a follow-up protocol of combined screening with sonography and mammography in an academic medical center. Other limitations are that our study population was composed of only Asian women and that the study was conducted at a single site with only a limited number of radiologists who perform a high annual volume of breast sonographic examinations.
Further studies of other ethnic groups and with radiologists of varied levels of experience are needed before we can generalize our results. However, importantly, recent reports have shown that consistently performing breast sonographic examinations and interpretations is possible with minimal training [28] and that breast cancer screening with combined sonography and mammography results in increased detection of early breast cancer, as seen in a study in which 93% of the study population were white women [13].
Finally, our study included the results not only from sonography but also from mammography. We suggest the usefulness of adjunctive sonography added to mammographic screening in high-risk patients. Our study population had a personal history of breast cancer surgery, one of the high-risk factors of breast cancer, and that characteristic of our population could have affected the results. Hence, a large study is mandatory to assess risk in patients of otherwise average risk who have dense breasts.
In conclusion, adding sonography to mammography for breast cancer screening of women with dense breasts and a history of surgery for breast cancer can help detect metachronous contralateral breast cancers predominantly in an early stage. Furthermore, considering the 0.06% false-negative result and the 41.0% PPV of the biopsy recommendation rate, our results suggest that sonography could be a useful adjunctive tool to mammography for the detection of metachronous contralateral breast cancers.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  H. J. Moon, M. J. Kim, E.-K. Kim, B.-W. Park, J. H. Youk, J. Y. Kwak, J. Sohn, and S.-I. Kim US Surveillance of Regional Lymph Node Recurrence after Breast Cancer Surgery Radiology, September 1, 2009; 252(3): 673 - 681. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
